{"id":13443923,"url":"https://github.com/jianzhnie/pyramidbox_pytorch","last_synced_at":"2025-09-06T13:32:12.093Z","repository":{"id":103457917,"uuid":"222218043","full_name":"jianzhnie/pyramidbox_pytorch","owner":"jianzhnie","description":"pytorch实现的Pyramidbox 人脸检测模型， 对原来代码的部分模块进行了修改，更简洁高效","archived":false,"fork":false,"pushed_at":"2020-12-08T11:58:35.000Z","size":77239,"stargazers_count":22,"open_issues_count":1,"forks_count":6,"subscribers_count":5,"default_branch":"master","last_synced_at":"2024-12-29T13:45:22.892Z","etag":null,"topics":["facedetection","pyramidbox","pytorch","ssd"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/jianzhnie.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null}},"created_at":"2019-11-17T08:23:53.000Z","updated_at":"2024-07-22T08:54:19.000Z","dependencies_parsed_at":"2024-01-18T15:24:57.602Z","dependency_job_id":null,"html_url":"https://github.com/jianzhnie/pyramidbox_pytorch","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jianzhnie%2Fpyramidbox_pytorch","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jianzhnie%2Fpyramidbox_pytorch/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jianzhnie%2Fpyramidbox_pytorch/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jianzhnie%2Fpyramidbox_pytorch/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/jianzhnie","download_url":"https://codeload.github.com/jianzhnie/pyramidbox_pytorch/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":232126160,"owners_count":18476190,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["facedetection","pyramidbox","pytorch","ssd"],"created_at":"2024-07-31T03:02:13.996Z","updated_at":"2025-01-01T20:56:00.584Z","avatar_url":"https://github.com/jianzhnie.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"## A PyTorch Implementation of PyramidBox ##\r\n[PyramidBox: A Context-assisted Single Shot Face Detector.](https://arxiv.org/abs/1803.07737)\r\n\r\n\r\n### Description\r\nI train pyramidbox with pytorch and the result approaches the original paper result,the pretrained model can be downloaded in [vgg](https://pan.baidu.com/s/1Q-YqoxJyqvln6KTcIck1tQ),the final model can be downloaded in [Pyramidbox](https://pan.baidu.com/s/1VtzgB9srkJY4SUtVM3n8tw).the AP in WIDER \r\nFACE as following:\r\n\r\n|             | Easy MAP | Medium MAP |  hard MAP |\r\n| --------    | ---------|------------| --------- |\r\n| origin paper|\t0.960    |    0.948   |  0.888    |\r\n| this repo   | 0.948    |    0.938   |  0.880    |\r\n\r\nthe AP in AFW,PASCAL,FDDB as following:\r\n\r\n| \tAFW     |   PASCAL\t|   FDDB   |\r\n| --------- |-----------| ---------|\r\n|\t99.65   |    99.02  |  0.983   |\r\n\r\nthe gap is small with origin paper,I train 120k batch_size 4 which is different from paper,which maybe cause the gap,if you have more gpu ,the final result maybe better.\r\n\r\n### Requirement\r\n* pytorch 1.3 \r\n* opencv \r\n* numpy \r\n* easydict\r\n\r\n### Prepare data \r\n1. download WIDER face dataset\r\n2. modify data/config.py \r\n3. ``` python tools/prepare_wider_data.py```\r\n\r\n\r\n### Train \r\n``` \r\npython main.py --batch_size 4  \r\n\t\t--lr 5e-4\r\n```\r\n\r\n### Evalution\r\naccording to yourself dataset path,modify data/config.py \r\n1. Evaluate on AFW.\r\n```\r\npython tools/afw_test.py\r\n```\r\n2. Evaluate on FDDB \r\n```\r\npython tools/fddb_test.py\r\n```\r\n3. Evaluate on PASCAL  face \r\n``` \r\npython tools/pascal_test.py\r\n```\r\n4. test on WIDER FACE \r\n```\r\npython tools/wider_test.py\r\n```\r\n### Demo \r\nyou can test yourself image\r\n```\r\npython demo.py\r\n```\r\n\r\n### Result\r\n\u003cdiv align=\"center\"\u003e\r\n\u003cimg src=\"https://github.com/jianzhnie/pyramidbox_pytorch/blob/master/data/img/gsmarena_001.jpg\" height=\"250px\" alt=\"demo\" \u003e\r\n\u003cimg src=\"https://github.com/jianzhnie/pyramidbox_pytorch/blob/master/data/img/0_Parade_marchingband_1_488.jpg\" height=\"200px\" alt=\"demo\" \u003e\r\n\u003c/div\u003e\r\n\r\n### References\r\n* [PyramidBox: A Context-assisted Single Shot Face Detector](https://arxiv.org/abs/1803.07737)\r\n* [PyramidBox model](https://github.com/PaddlePaddle/models/tree/develop/fluid/PaddleCV/face_detection)\r\n* [S3FD_pytorch](https://github.com/jianzhnie/S3FD_pytorch)\r\n* [ssd.pytorch](https://github.com/amdegroot/ssd.pytorch)\r\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjianzhnie%2Fpyramidbox_pytorch","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjianzhnie%2Fpyramidbox_pytorch","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjianzhnie%2Fpyramidbox_pytorch/lists"}